On this page:
pydrnlp:   NLP Library for Digital Ricœur

pydrnlp: NLP Library for Digital Ricœur

Philip McGrath <philip@philipmcgrath.com>

 (require pydrnlp) package: pydrnlp

The pydrnlp repository contains:
  1. A Python 3 package providing some NLP functionality needed by Digital Ricœur, which is implemented using spaCy; and

  2. A Racket package that manages the Python-implemented functionality and integrates it with other Digital Ricoeur tools.

The initial goal is to support the NLP functionality needed to implement our own versions of some of the widgets that Digital Ricœur currently gets from Voyant.

The functionality described in this document is still at an experimental stage.

To use the Python functionality, you must have a conda executable in your PATH for the Conda package manager. The recommended way to obtain this is by installing the Miniconda distribution, though the larger Anaconda distribution should work as well. As long as you have conda installed, raco setup will manage the Python dependencies and virtual environment automatically. Note that you do not need to have Conda or even Python installed merely to build pydrnlp (including the Python documentation) or to use the Racket-implemented functionality.

NOTE: Much of this document is badly out-of-date. We are in the process of updating it in conjunction with the initial release.

    1 “Trends” Tool

      1.1 Implementation Details

        1.1.1 Python Layer

    2 Framework for Implementing Additional Tools

      2.1 Managing Python Processes from Racket

      2.2 Python–Racket Bridge Language

      2.3 Scribbling Python Documentation

      2.4 Python Utility Modules

        2.4.1 pydrnlp.language

        2.4.2 pydrnlp.jsonio